`cluster_dump` roundtrips break non-stringified keys

[hendrikmakait]

Describe the issue:

After we removed key stringification in #8083, keys can now be composite of multiple different types. Our cluster_dump functionality is not capable of correctly roundtripping those.

Minimal Complete Verifiable Example:

import dask.dataframe as dd
import pandas as pd
from distributed import Client
from distributed.cluster_dump import DumpArtefact

if __name__ == "__main__":
    c = Client()
    df = dd.from_pandas(pd.DataFrame({"a": [1, 2, 3]}), npartitions=3)
    result = df.compute()
    c.dump_cluster_state(filename="dump", format="yaml")
    dump = DumpArtefact.from_url("dump.yaml")
    dump.scheduler_story(df.__dask_keys__()[0])

raises

Traceback (most recent call last):
  File "/Users/hendrikmakait/projects/dask/distributed/foo.py", line 12, in <module>
    dump.scheduler_story(df.__dask_keys__()[0])
  File "/Users/hendrikmakait/projects/dask/distributed/distributed/cluster_dump.py", line 207, in scheduler_story
    for story in _scheduler_story(keys, log):
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/hendrikmakait/projects/dask/distributed/distributed/_stories.py", line 33, in scheduler_story
    if t[0] in keys_or_stimuli or keys_or_stimuli.intersection(t[3])
       ^^^^^^^^^^^^^^^^^^^^^^^
TypeError: unhashable type: 'list'

For msgpack, the error is the same.

[hendrikmakait]

In an offline conversation, @fjetter, @crusaderky, and I have decided not to put effort into this, as cluster dumps have rarely been useful in the last few months. If something similar becomes necessary again, we will likely return to the drawing board and rebuild this from the ground.

If someone finds this useful and wants to go through the effort of fixing cluster dumps in the meantime, feel free to do so.